inside /usr/local/cuda/include/vector_types.h, I see that cuda’s built-in vectors have special tags that I cannot find documentation to:
__device_builtin__
__align__(n)
__builtin_align__(n)
What do these three do?
inside /usr/local/cuda/include/vector_types.h, I see that cuda’s built-in vectors have special tags that I cannot find documentation to:
__device_builtin__
__align__(n)
__builtin_align__(n)
What do these three do?